Behavioral and physiological flexibility are used by birds to manage energy and support investment in the early stages of reproduction

Current Zoology 56 (6): 767 792, 2010 Behavioral and physiological flexibility are used by birds to manage energy and support investment in the early stages of reproduction François VÉZINA 1*, Katrina G. SALVANTE 2 1 Département de Biologie, chimie et géographie, Groupe de recherche sur les environnements nordiques BORÉAS, Université du Québec à Rimouski, 300 Allée des Ursulines, Rimouski, Québec, Canada, G5L 3A1 2 Department of Biological Sciences, Simon Fraser University, 8888 University Drive, Burnaby, BC, Canada, V5A 1S6 Abstract Interest in phenotypic flexibility has increased dramatically over the last decade, but flexibility during reproduction has received relatively little attention from avian scientists, despite its possible impact on fitness. Because most avian species maintain atrophied reproductive organs when not active, reproduction in birds requires major tissue remodeling in preparation for breeding. Females undergo rapid (days) recrudescence and regression of their reproductive organs at each breeding attempt, while males grow their organs ahead of time at a much slower rate (weeks) and may maintain them at maximal size throughout the breeding season. Reproduction is associated with significant metabolic costs. Egg production leads to a 22% 27% increase in resting metabolic rate (RMR) over non-reproductive values. This is partly due to the activity of the oviduct, an organ that may allow females to adjust reproductive investment by modulating egg size and quality. In males, gonadal recrudescence may lead to a 30% increase in RMR, but the data are inconsistent and general conclusions regarding energetic costs of reproduction in males will require more research. Recent studies on captive female zebra finches describe the impacts of these costs on daily energy budgets and highlight the strategies used by birds to maintain their investment in reproduction when energy is limited. Whenever possible, birds use behavioral flexibility as a first means of saving energy. Decreasing locomotor activity saves energy during challenges such as egg production or exposure to cold temperatures and is an efficient way to buffer variation in individual daily energy budgets. However, when behavioral flexibility is not possible, birds must rely on flexibility at the physiological level to meet energy demands. In zebra finches breeding in the cold, this results in a reduced pace of laying, likely due to down-regulation of both reproductive and non-reproductive function, allowing females to defend minimal egg size and maintain reproductive success. More research involving a range of species in captive and free-living conditions is needed to determine how phenotypic flexibility during tissue remodeling and early reproductive investment translates to natural conditions and affects fitness [Current Zoology 56 (6): 767 792, 2010]. Key words Phenotypic plasticity, Phenotypic flexibility, BMR, Energy budget, Organ, Egg size, Physiological tradeoff, Fitness 1 Introduction Phenotypic flexibility is defined as the capacity for an individual organism to reversibly transform its phenotype (Piersma and Drent, 2003). It is one of four subcategories of phenotypic plasticity that also include developmental plasticity, polyphenism and life-cycle staging (Piersma and Drent, 2003). Over the last two decades, evolutionary biologists have shown a growing interest in phenotypic flexibility. A search on Web of Science for papers specifically using the term phenotypic flexibility in their title, abstract or key words shows a clear trend for an increasing number of publications per year referring to the phenomenon (Fig. 1), with numbers taking off in years following Piersma and Drent s (2003) conceptual distinction of flexibility within plasticity. Although confusion remains with the use of the broader term phenotypic plasticity, this simple survey clearly shows that interest in phenotypic flexibility is growing in the collective scientific mind. Although all spheres of evolutionary biology may now include studies on adaptive phenotypic flexibility, avian research has been particularly active over the last decade. Indeed, of all papers published since 1990 that were reported by Web of Science as using the term phenotypic flexibility, 43% (71 of 167 papers) were based on or involved avian systems (using key word bird* ). There are many examples of phenotypic flexibility in birds (e.g. Piersma and Lindstom, 1997; Starck, 2005; McKechnie, 2008; Swanson, 2010) but some Received May 09, 2010; accepted Aug 23, 2010 Corresponding author. E-mail: francois_vezina@uqar.qc.ca 2010 Current Zoology

768 Current Zoology Vol. 56 No. 6 Fig. 1 Number of articles using the term phenotypic flexibility in their title, abstract or keywords since 1990 according to Web of Science The search used the Science Citation Index Expanded Database and excluded year 2010. aspects of their life history have received more attention than others. For example, long distance migration is generally recognized as a highly demanding activity involving considerable flexibility in several physiological traits, from endocrine adjustments (e.g. Holberton, 1999, Piersma et al., 2000; Landys-Ciannelli et al., 2002; Landys et al., 2004) to changes in nutrient transport and fuel use (McWilliams et al., 2004, Weber, 2009, Swanson, 2010), organ size (e.g. Piersma et al., 1996, Biebach, 1998, Battley et al., 2000, Piersma, 2002; Guglielmo and Williams, 2003; Bauchinger et al., 2005) and metabolic performance (Swanson and Dean, 1999; Battley et al., 2001, Kvist and Lindstrom, 2001; McKechnie, 2008). Birds have also been shown to exhibit remarkable flexibility at several levels of integration when faced with specific food types requiring reversible adjustments in their digestive machinery (Piersma et al., 1993; Piersma et al., 1999; Starck, 1999; Dekinga et al., 2001; Battley and Piersma, 2005). Flexibility in metabolic performance (measured as basal metabolic rate; BMR, and maximal thermogenic capacity; M sum ), evaporative water loss and skin structure, as well as in the size of internal organs have also been highlighted in the context of seasonal acclimatization and laboratory acclimation to different temperatures (Williams and Tieleman, 2000; Haugen et al., 2003; Tieleman et al., 200; McKechnie et al., 2007; Cavieres and Sabat, 2008; McKechnie, 2008; Zheng et al., 2008, Barcelo et al., 2009; Swanson, 2010). In this specific case, although the majority of studies have focused on non-migratory bird species from northern latitudes (Smit and McKechnie, 2010), new evidence from studies including migratory species or species experiencing seasonally stable, relatively mild or dry environments support the hypothesis that phenotypic flexibility in response to thermal regime is a common feature of birds (Williams and Tieleman, 2000; Tieleman et al., 2003; Klaassen et al., 2004; McKechnie et al., 2007; Cavieres and Sabat, 2008; Barcelo et al., 2009; Maldonado et al., 2009; Salvante et al., 2010). With the realization that a wide variety of phenotypic traits in animals are flexible (Piersma and Drent, 2003), scientists are now interested in testing the adaptive value of phenotypic flexibility by linking flexibility with fitness (e.g. Ricklefs and Wikelski, 2002; Seebacher, 2005; Naya et al., 2008; Moore and Hopkins, 2009). A surprising fact associated with the growing interest in phenotypic flexibility is that perhaps one of the best-known examples of avian physiological flexibility appears to have attracted very little attention. According to Web of Science, of the 71 papers on avian systems published since 1990 which use the term phenotypic flexibility, only one (1.4%) explicitly studied flexibility in reproductive traits (Partecke et al., 2004). However, this rough survey misses studies wrongly referring to plasticity instead of flexibility. The same analysis using the term phenotypic plasticity shows that only 35% of studies investigated aspects of reproduction. Yet, reproduction in birds is associated with major phenotypic changes at multiple levels, from endocrine induction to internal organ reorganization leading to new tissue synthesis (e.g., reproductive organ growth, egg production) (Opel and Nalbanov, 1961a, b;yu et al., 1971; Follett and Maung, 1978; Dawson and Goldsmith, 1983; Burley and Vadehra, 1989; Etches, 1996; Williams, 1998; Vézina and Williams, 2003). These transformations have measurable energetic costs (Chappell et al., 1999; Nilsson and Raberg, 2001; Vézina and Williams, 2002, 2005a; Salvante et al., 2010), and these costs may result in reversible changes in behavior and physiology with consequences for energy budgets (Houston et al., 1995; Williams and Ternan, 1999; Vézina and Williams, 2005b; Vézina et al., 2006; Salvante et al., 2007, 2010; Williams et al., 2009). The scarcity of avian studies on phenotypic flexibility during reproduction is even more surprising given that phenotypic changes associated with variation in individual reproductive investment may presumably be directly linked with reproductive success (e.g. oviduct size vs egg size and quality, Ricklefs 1976, Christians and Williams 2001a, Vézina and Williams 2003), a major component of fitness. Physiological reproductive effort (i.e. egg production)

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 769 has also been linked to adult survival, the second component of fitness. Indeed manipulative experiments have shown negative effects of experimentally increased reproductive effort on reappearance of adult females the following years (Heaney and Monaghan, 1995; Nager et al., 2001; Visser and Lessells, 2001; Nager, 2006). These effects are likely mediated by long-term consequences on physiological condition (reviewed by Nager, 2006). Similar findings in natural populations experiencing mismatches between the timing of reproduction and the natural peak of food resources also suggest a survival cost to reproductive effort (e.g. Thomas et al., 2001). With this overview, our aim is to highlight current information on avian reproductive phenotypic flexibility in the context of energy management strategies. We suggest new research avenues that are likely to help in understanding how flexibility may influence reproductive success by allowing breeding birds to balance their energy budgets. The notion that reproduction is a highly energy-demanding life-history stage in birds has been formulated from decades of studies on avian reproduction (e.g. Drent and Daan, 1980) which, for the most part, have focused on the periods of incubation or, in altricial species, nestling provisioning (Williams and Vézina, 2001; Williams, 2005; Nager, 2006). These life-history stages certainly deserve our attention in the context of adaptive phenotypic flexibility and fitness. However this paper will focus on the much less studied period of reproductive organ recrudescence, egg formation and laying. 2 Seasonal Remodeling of Body Composition Timing reproductive effort to coincide with resource availability in a seasonal environment is crucial for birds as it may significantly contribute to reproductive success and fitness. Reproduction in birds is generally timed to match the peak of nestling demand with maximal food abundance or the seasonal availability of specific dietary resources (Perrins, 1996; Williams, 1998). However, the vast majority of avian species maintain their reproductive organs in an atrophied state throughout all non-reproductive stages (Williams, 1998). Consequently, gonadal recrudescence, egg formation and fertilization, egg laying and incubation, which all occur prior to the period of nestling care, generally happen at a time when environmental conditions may be suboptimal in terms of nutrient and energy availability (Perrins, 1996). Yet major reversible organ changes occur in both sexes in preparation for breeding. With global climate change, mismatches in timing of breeding and food resource availability (e.g. Visser et al., 2006) are likely to exacerbate this situation and further reduce energy availability for pre-breeding body remodeling. In females, reported changes in oviduct size from a non-breeding to a reproductively mature state can range between 5- to 220-fold, while ovarian development can lead to a mature organ that is 4- to 75-fold larger than the immature state (Table 1). In males, testicular growth exhibits a larger magnitude of increase, with mature testes growing up to 480-fold larger than their non-reproductive size (Table 2). However, mature reproductive organs of females typically represent a larger proportion of total body mass compared to those of males (average 8.6% and 1.5% of non-reproductive body mass in females and males, respectively, considering only wet mass data and natural conditions from Tables 1 and 2). In females of species where the pattern of organ recrudescence has been detailed, growth of reproductive organs happens in synchrony with rapid yolk development, preceding the ovulation of the first ovarian follicle. This period can be very rapid in small species (4 days in the opportunistic breeding zebra finch Taenyopigia guttata, 6 days in the seasonal breeding European starling, Sturnus vulgaris) and may take up to 25 days in larger species like the Pacific black brant Branta bernicla nigricans (Table 1). In males, however, gonadal recrudescence is a much slower process that can take 2 8 weeks to complete under natural conditions (Table 2), often beginning as early as during the winter season (e.g. Hegner and Wingfield, 1986c; Wikelski et al., 2003; Raes and Gwinner, 2005; Caro et al., 2006). Even though several bird species lay more than one clutch per season, recent evidence in passerines strongly suggests that individual females go through cycles of recrudescence and regression of their reproductive organs at each reproductive attempt in order to minimize the energetic costs associated with maintaining and carrying these organs (Vézina and Williams, 2003; Williams and Ames, 2004). In contrast, males can maintain their mature testes for much longer (Table 2), up to six months in rufous-winged sparrows Aimophila carpalis (Small et al., 2007). However, the pattern of recrudescence and regression may differ significantly among species and breeding strategies. For instance, female house sparrows Passer domesticus produce several clutches of eggs per breeding season and male testes remain fully developed for the entire breeding season

770 Current Zoology Vol. 56 No. 6 Table 1 Gonadal development in female birds Species Reproductive organ/tissue ANRBM NRSO RSO XIS RRM MDG MDM Measurement conditions References Pacific black brant Ovary mass (g)* 1431* 0.4 2.15 5.375 0.15 < 25 days Natural Mason et al., 2007 Branta bernicla nigricans Mallard duck Oviduct mass (g) W 967.3* 7.2 32.1 4.5 3.32 Natural Krapu, 1981 Anas platyrhynchos Ovary mass (g) W 1.8 31.9 17.7 3.30 Natural Krapu, 1981 Ruddy duck Oviduct mass (g) D 529 0.9 7.7 8.6 1.46 Natural Oxyura jamaicensis Ovary mass (g) D 0.2 15.04 75.2 0.28 Natural Tome, 1984; mass: Hohman et al., 1992 Tome, 1984; mass: Hohman et al., 1992 European barn owl Ovary mass (mg)* 305.8 30.9 182.0 5.9 0.06 Tyto alba Ovary mass (mg)* 39.8 199.5 5.0 0.07 Ovary mass (mg)* 25.1 398.1 15.8 0.13 of birds with fat score >1 of adult birds of first year birds mass: Massemin et al., 1997 mass: Massemin et al., 1997 mass: Massemin et al., 1997 European sparrowhawk Ovary mass (mg)* 260* 15.1 186.2 12.3 0.07 of birds with fat score >1 mass: Vedder et al., 2005 Accipiter nisus Ovary mass (mg)* 31.6 251.2 7.9 0.10 Ovary mass (mg)* 13.8 177.8 12.9 0.07 Kestrel Ovary mass (mg)* 217 31.6 316.2 10.0 0.15 Falco tinnunculus Ovary mass (mg)* 39.8 316.2 7.9 0.15 Ovary mass (mg)* 12.6 158.5 12.6 0.07 of adult birds of first year birds of birds with fat score >1 of adult birds of first year birds mass: Vedder et al., 2005 mass: Vedder et al., 2005 mass: Jonsson et al., 1996 mass: Jonsson et al., 1996 mass: Jonsson et al., 1996 European starling Oviduct mass (g) LD 77 0.014 0.965 68.9 1.25 6 days Natural Vézina unpublished data (to be continued on the next page)

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 771 Table 1 (Continued) Species Reproductive organ/tissue ANRBM NRSO RS XIS RRM MDG MDM Measurement conditions References Sturnus vulgaris Ovary mass (g) LD 0.01 0.05 5.0 0.06 6 days Natural Vézina unpublished data Ovarian follicle mass (g) LD 0.01 0.28 28.0 0.36 6 days Natural Vézina unpublished data Oviduct mass (g) W 0.08 3.99 49.9 5.18 6 days Natural Vézina unpublished data Ovary mass (g) W 0.07 0.27 3.9 0.35 6 days Natural Vézina unpublished data Clay-colored robin Turdus grayii Blue-gray tanager Thraupis episcopus Bicolored antbird Ovarian follicle mass (g) W Largest ovarian follicle diameter (mm)* Largest ovarian follicle diameter (mm)* Largest ovarian follicle diameter (mm)* 0.01 1.19 119.0 1.55 6 days Natural Vézina unpublished data 67.9* 0.5 4.7 9.4 Natural Wikelski et al., 2003 32.8* 0.5 3.7 7.4 Natural Wikelski et al., 2003 29.4* 0.5 3.1 6.2 Natural Wikelski et al., 2003 Gymnopithys leucaspis bicolor House sparrow Oviduct mass (mg) D 28.4 0.006 0.333 55.5 1.17 Krementz and Ankney, 1986; mass: Johnston and Selander, 1973 Passer domesticus House sparrow Largest ovarian follicle diameter (mm)* 27.5* 0.1 6 60 <3 weeks Natural Hegner and Wingfield, 1986b Passer domesticus Pacific black brant Ovary mass (g)* 1431* 0.4 2.15 5.375 0.15 < 25 days Natural Mason et al., 2007 Branta bernicla nigricans Mallard duck Oviduct mass (g) W 967.3* 7.2 32.1 4.5 3.32 Natural Krapu, 1981 (to be continued on the next page)

772 Current Zoology Vol. 56 No. 6 Table 1 (Continued) Species Reproductive organ/tissue ANRBM NRSO RS XIS RRM MDG MDM Measurement conditions References Anas platyrhynchos Ovary mass (g) W 1.8 31.9 17.7 3.30 Natural Krapu, 1981 Ruddy duck Oviduct mass (g) D 529 0.9 7.7 8.6 1.46 Natural Oxyura jamaicensis Ovary mass (g) D 0.2 15.04 75.2 0.28 Natural Tome, 1984; mass: Hohman et al., 1992 Tome 1984; mass: Hohman et al. 1992 European barn owl Ovary mass (mg)* 305.8 30.9 182.0 5.9 0.06 Tyto alba Ovary mass (mg)* 39.8 199.5 5.0 0.07 Ovary mass (mg)* 25.1 398.1 15.8 0.13 European sparrowhawk Ovary mass (mg)* 260* 15.1 186.2 12.3 0.07 Accipiter nisus Ovary mass (mg)* 31.6 251.2 7.9 0.10 of birds with fat score >1 of adult birds of first year birds of birds with fat score >1 of adult birds mass: Massemin et al., 1997 mass: Massemin et al., 1997 mass: Massemin et al., 1997 mass: Vedder et al., 2005 mass: Vedder et al., 2005 Song wren Largest ovarian follicle diameter (mm)* 23* 0.5 2.3 4.6 Natural Wikelski et al., 2003 Cyphorhinus phaeocephalus White-crowned sparrow Zonotrichia leucophrys gambelii Ovary mass (mg) W F 23 4.9 18.6 3.8 0.08 Natural: Spring migration Farner et al., 1966 Ovary mass (mg) W F 4.9 326 66.5 1.42 Natural: Breeding maximum Farner et al., 1966 Ovarian follicle diameter (mm)* 0.5 > 5 at least 10 Natural: Categorized: non-breeding <0.5 mm, reproductive >5 mm Wingfield and Farner, 1978a Zonotrichia leucophrys pugetensis Ovarian follicle diameter (mm)* 24.5* 0.5 > 5 at least 10 ~1 week ~7 days Natural: Categorized: non-breeding <0.5 mm, reproductive >5 mm Wingfield Farner, 1978b (to be continued on the next page)

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 773 Table 1 (Continued) Species Reproductive organ/tissue ANRBM NRSO RS XIS RRM MDG MDM Measurement conditions References Great tit Oviduct mass (mg)* W F 18.34 1 220 220 1.20 Natural Silverin, 1978; mass: Broggi et al., 2007 Parus major Ovary mass (mg)* W F 1 65 65 0.35 Natural Zebra finch Oviduct mass (mg) LD 17 5.97 120 20.1 0.71 4 days Taeniopygia guttata < 10 days Silverin, 1978; mass: Broggi et al., 2007 Natural: Captive Houston et al., 1995 Spotted antibird Hylophylax n. naevioides Golden-collared manakin Manacus vitelinus Red-capped manakin Pipra mentalis Largest ovarian follicle diameter (mm)* Largest ovarian follicle diameter (mm)* Largest ovarian follicle diameter (mm)* 16.9* 0.5 7.5 15 Natural Wikelski et al., 2003 15.5* 0.5 4.7 9.4 Natural Wikelski et al., 2003 13.8* 0.5 6 12 Natural Wikelski et al., 2003 Pied Flycatcher Oviduct mass (g) LD 12.30* 0.04 0.24 6 1.95 Natural Ojanen, 1983b Ficedula hypoleuca Ovary mass (g) LD 0.01 0.12 12 0.98 Natural Ojanen, 1983b Oviduct mass (g) W 0.18 1.00 5.6 8.13 Natural Ojanen, 1983b Ovary mass (g) W 0.04 0.67 16.75 5.45 Natural Ojanen, 1983b Williow tit Largest ovarian follicle diameter (mm)* 10.94 0.25 0.96 3.84 Experimental (may not be fully developed) Silverin and Westin, 1995; mass: Broggi et al., 2003 Parus montanus All ovary masses do not include the mass of the ovarian follicles. * Approximated from figure. D: Dry. F: Fixed. LD: Lean dry. W: Wet. ANRBM: Average non-reproductive body mass (g). MDG: Minimum duration of organ/tissue growth. MDM: Minimum duration of maintenance of fully-developed organ/tissue. NRSO: Non-reproductive size of organ/tissue. RRM: Relative reproductive organ/tissue mass (compared to non-reproductive body mass) (%). RSO: Reproductive size of organ/tissue. XIS: X-fold increase in organ/tissue size relative to its non-reproductive size.

774 Current Zoology Vol. 56 No. 6 Table 2 Gonadal development in male birds Species Reproductive organ/tissue ANRBM NRSO RSO XIS RRM MDG MDM Measurement conditions References Pacific black brant Combined testes mass (g)* 1610* 0.4 2.4 6.0 0.15 < 40 days Natural Mason et al., 2007 Branta bernicla nigricans European barn owl Combined testes mass (mg)* 286.9 32.4 295.8 9.1 0.10 few months Tyto alba Combined testes mass (mg)* 41.8 490.9 11.7 0.17 few months Combined testes mass (mg)* 32.4 257.6 7.9 0.09 few months Black-billed magpie Combined testes mass (mg) 186.5 4 1250 312.5 0.67 ~3 months Natural Pica pica Kestrel Combined testes mass (mg)* 185 10.5 224.4 21.4 0.12 few months Falco tinnunculus Combined testes mass (mg)* 21.4 229.6 10.7 0.12 few months Combined testes mass (mg)* 6.3 126.2 20.0 0.07 few months European sparrowhawk Combined testes mass (mg)* 160* 10.5 200.0 19.1 0.13 few months Accipiter nisus Combined testes mass (mg)* 11.5 257.6 22.4 0.16 few months Combined testes mass (mg)* 10.7 135.2 12.6 0.08 few months of birds with fat score >1 of adult birds of first year birds of birds with fat score >1 of adult birds of first year birds of birds with fat score >1 of adult birds of first year birds European starling Combined testes mass (mg)* W 81.5 4 800 200.0 0.98 4 weeks Experimental Sturnus vulgaris Combined testes mass (mg)* W 4 1750 437.5 2.15 ~60 days 1.5 2 months Natural Testis volume (mm 3 )* 2.5 530.0 212.0 ~2 months Experimental Testis volume (mm 3 )* 2.5 500.0 200.0 ~2 months Natural mass: Massemin et al., 1997 Young et al. 2009, mass: Massemin et al., 1997 mass: Massemin et al., 1997 Erpino, 1969; Mass: Trost, 1999 mass: Jonsson et al., 1996 mass: Jonsson et al. 1996 mass: Jonsson et al., 1996 mass: Vedder et al., 2005 mass: Vedder et al., 2005 mass: Vedder et al., 2005 Dawson et al., 2002; mass: Hicks, 1934 Ball and Ketterson, 2008; mass: Hicks, 1934 Dawson, 2005; mass: Hicks, 1934 Dawson, 2005; mass: Hicks, 1934 (to be continued on the next page)

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 775 Species Reproductive organ/tissue ANRBM NRSO RS XIS RRM MDG MDM Measurement conditions References Groove-billed ani Combined testes mass (mg) W 79 Crotophaga sulcirostris Testis length (mm) Combined testes mass (mg) W 73.5 Testis length (mm) 54 (midincubation) 4.5 (midincubation) 202 (midincubation) 7.5 (midincubation) 380 (laying) 7.0 0.48 10 (laying) 2.2 244 (earlyincubation) 9 (earlyincubation) 1.2 0.33 1.2 Natural: Alpha male = primary incubator Natural: Alpha male = primary incubator Natural: Subordinate males = limited incubation behaviour Natural: Subordinate males = limited incubation behaviour Table 2 (Continued) Vehrencamp, 1982 Vehrencamp, 1982 Vehrencamp, 1982 Vehrencamp, 1982 Clay-colored robin Testis volume (mm 3 )* 75.7* 0.5 8.7 17.4 Natural Wikelski et al., 2003 Turdus grayii Brown-headed cowbird **Estimated combined testes mass (mg) 47* 50 780 15.6 1.66 4 weeks ~6 weeks Natural Dufty and Wingfield, 1986 Molothrus ater Blue-gray tanager Testis volume (mm 3 )* 34.2* 0.5 5.3 10.6 Natural Wikelski et al., 2003 Thraupis episcopus Bicolored antbird Testis volume (mm 3 )* 30.7* 1.0 7.1 7.1 Natural Wikelski et al., 2003 Gymnopithys leucaspis bicolor House sparrow **Estimated combined testes mass (mg)* 27.5* 12 1000 83.3 3.64 4 months Passer domesticus Natural (not sure if non-reproductive testis size is basal) Hegner Wingfield, 1986a **Estimated combined testes mass (mg)* 4 1000 250.0 3.64 ~2 months Natural Hegner Wingfield, 1986c White-crowned sparrows Combined testes mass (mg) W F 25.5 2.2 95.4 43.4 0.37 Natural: Spring migration Farner et al., 1966 Zonotrichia leucophrys Combined testes mass (mg) W F gambelii 2.2 884.0 401.8 3.47 Natural: Breeding max Farner et al., 1966 **Estimated combined testes mass (mg)* 25.5 2 960 480.0 3.76 ~4 weeks ~4 6 weeks Natural Wingfield Farner, 1978a Zonotrichia leucophrys **Estimated combined testes mass (mg)* pugetensis 26.5* 2 1000 500.0 3.77 ~2 weeks ~3 months Natural Wingfield Farner, 1978b Song wren Testis volume (mm 3 )* 25.4* 0.5 5.7 11.4 Natural Wikelski et al., 2003 Cyphorhinus phaeocephalus Great tit Testis length (mm)* 18.34 1.25 6.5 7 5.2 5.6 15 43 days Experimental Silverin et al., 2008; mass: Broggi et al., 2007 (to be continued on the next page)

776 Current Zoology Vol. 56 No. 6 Species Reproductive organ/tissue ANRBM NRSO RS XIS RRM MDG MDM Measurement conditions References Parus major Table 2 (Continued) Testis volume (mm 3 )* 1.0 2.5 2.75 2.5 2.75 11 weeks Experimental Combined testes mass (mg)* W F 2 260 130.0 1.42 Natural Caro and Visser, 2009; mass: Broggi et al., 2007 Silverin, 1978; mass: Broggi et al., 2007 Golden-collared manakin Testis volume (mm 3 )* 18.1* 0.5 9.3 18.6 Natural Wikelski et al., 2003 Manacus vitelinus Spotted antbird Testis volume (mm 3 )* 17.8* 3.5 11.3 3.2 6 weeks Hylophylax n. naevioides Natural (not sure if non-reproductive testis size is basal) Hau et al., 2000 Testis volume (mm 3 )* 17* 0.5 9.6 19.2 Natural Wikelski et al., 2003 Dark-eyed junco Combined testes mass (mg)* W 17.6* 5 384 76.8 2.18 ~5 weeks 2 months Natural: After second year males Deviche et al., 2000 Junco hyemalis Combined testes mass (mg)* W 5 339 67.8 1.93 ~5 weeks 2 months Natural: Second year males Deviche et al., 2000 Paired testes volume (mm 3 ) 1.2 116.2 96.8 4 months European stonechats Testis width (mm)* 16 1 4.5 4.5 Saxicola torquata rubicola Experimental (not sure if testes are fully developed) Natural (not sure if non-reproductive testis size is basal) Engels and Jenner, 1956 Raess and Gwinner, 2005; mass: Klaassen, 1995 Siberian stonechats Testis width (mm)* 0.4 4.8 12 Natural Raess and Gwinner, 2005 Saxicola torquata maura Rufous-winged sparrows Testis diameter (mm)* 15.5* 1.0 4.5 4.5 6 months Natural Small et al., 2007 Aimophila carpalis Red-capped manakin Testis volume (mm 3 )* 14* 0.5 8.7 17.4 Natural Wikelski et al., 2003 Pipra mentalis (to be continued on the next page)

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 777 Species Reproductive organ/tissue ANRBM NRSO RSO XIS RRM MDG MDM Measurement conditions References Table 2 (Continued) Pied Flycatcher Combined testes mass (mg)* W F 12.1 2.93 73.34 25.0 0.61 Natural Silverin, 1975; mass: Saetre et al., 1995 Ficedula hypoleuca Williow tit Testis length (mm)* 11.53 1.3 5.0 3.8 15 days Parus montanus Experimental (not sure if testes are fully developed) Silverin and Westin, 1995; mass: Broggi et al., 2003 * Approximated from figure. ** Based on comparisons with preserved specimens of known mass D: Dry. F: Fixed. LD: Lean dry. W: Wet. ANRBM: Average non-reproductive body mass (g). MDG: Minimum duration of organ/tissue growth. MDM: Minimum duration of maintenance of fully-developed organ/tissue. NRSO: Non-reproductive size of organ/tissue. RRM: Relative reproductive organ/tissue mass (compared to non-reproductive body mass) (%). RSO: Reproductive size of organ/tissue. XIS: X-fold increase in organ/tissue size relative to its non-reproductive size.

778 Current Zoology Vol. 56 No. 6 (Hegner and Wingfield, 1986b). In single-brooded species however, males may regress their testes rapidly following a breeding attempt. This is the case in pied flycatchers Ficedula hypoleuca from Sweden where a sharp decline in testes mass is observed after clutch completion (Silverin, 1975). In this specific case, males are breeding once per season, and females that lose their eggs can re-lay a new clutch, but only when paired with a different male. In single-brooded populations of white-crowned sparrows Zonotrichia leucophrys, testes begin regressing during incubation and decline rapidly during nestling provisioning (Wingfield and Farner, 1978a), but in populations where breeding pairs produce a second brood, the testes regress only slightly (20%) during incubation and grow back during the formation of the second clutch of eggs (Wingfield and Farner, 1978b). In the groove-billed ani Crotophaga sulcirostris, a tropical, communally-nesting cuckoo, the pattern of testes recrudescence and regression depends on the male s social status (Vehrencamp, 1982). Dominant males are the primary incubators, and they significantly regress their testes during incubation. Meanwhile the magnitude of testes regression in subordinate males, which perform the least amount of incubation, is much less pronounced. Anis are opportunistic breeders and breed during most of the year. Maintaining large, functional testes throughout the breeding season presumably allows for multiple breeding events and extra pair copulation. However, this may be incompatible with male incubation behavior. Non-reproductive organs have also been reported to change in association with gonadal recrudescence but the pattern of organ flexibility appears inconsistent, at least in females, among species and among years within species (Christians and Williams, 1999; Vézina and Williams, 2003; Nager, 2006). These changes will be further discussed in the sections below. 3 The Metabolic Cost of Reproductive Investment 3.1 Females The energetic investment in egg production has now been measured empirically five times in four passerine species (Table 3). The best estimates come from comparisons of resting or basal metabolic rates (RMR and BMR, respectively) in females measured before gonad recrudescence and during active egg production (i.e. in females with fully mature reproductive organs and an egg in the oviduct). Note that the term basal metabolic rate technically refers to animals not actively involved in new tissue synthesis. Although this criterion may be met in non-reproductive birds, it is not the case in birds measured during egg production and active spermatogenesis. While some authors did not make this distinction, for clarity we use the term resting metabolic rate hereafter. Measurements in free living great tits Parus major, European starlings and captive zebra finches showed that females actively producing eggs have a resting metabolic rate 22% 27% above non-reproductive RMR values. Estimates are consistent for the three species whether measurements were performed within or among individuals (Table 3), and this cost has been shown to be repeatable in zebra finches (Vézina and Williams, 2005a). The cost of reproductive investment in females has also been measured in house sparrows using a different approach (Chappell et al., 1999). All birds were measured during the reproductive period and, although Table 3 Energy cost of egg production in species for which measurements have been made by comparing resting or basal metabolic rates of females before gonad recrudescence and during active egg production Species Reference stage Breeding stage % increase in RMR 5 Comparison type Reference House sparrow Passer domesticus Great tit Parus major European starling Sturnus vulgaris Zebra finch Taeniopygia guttata Zebra finch Taeniopygia guttata Non breeding 1 Large gonads 3 16 Among individuals Chappell et al.,1999 Wintering 1 6 eggs laid (27) Among individuals Nilsson and Raberg, 2001 Non breeding 2 Six yolky follicle 4 22 Among individuals Vézina and Williams, 2002 Non breeding First egg laid 22 Within individuals Vézina and Williams, 2005 Non breeding First egg laid 24 (34) Within individuals Salvante et al., 2010 1 Females caught during the breeding season with gonads size 1/50 of mean gonad mass. 2 Females caught in the spring before gonads recrudescence. 3 Five time mean gonadal mass. 4 Day preceding first ovulation. 5 % increase in RMR above non-reproductive RMR values. Based on data corrected for mass by regression or ANCOVA. Values in brackets are based on estimates with no mass correction.

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 779 no measured females had begun laying (no eggs observed in the oviduct, M.A. Chappell pers. comm.), some had large yolky follicles (i.e., gonadal tissue mass five times the mean gonadal tissue mass of all measured females). Comparing predicted RMR estimates (based on a multiple regression predicting log RMR from log body mass and dry gonadal mass) for females with actively yolking ovarian follicles with those of females exhibiting fully regressed reproductive organs showed a 16% higher RMR in birds with large reproductive organs. This estimate is lower than measurements reported above for birds that were actively laying, but it is consistent with Nilsson and Raberg s (2001) observation of a 12% higher RMR in female great tits during ovarian recrudescence. Therefore, RMR appears to be increasing gradually during rapid yolk development and reproductive organ growth to culminate when the bird has an egg in its fully-grown oviduct (but see Vézina and Williams, 2002). A third approach, based on daily energy expenditure (DEE) measurements has also been used to investigate energy investment in egg production. Ward and MacLoed (1992) used respirometry to measure DEE in Japanese quails (Coturnix coturnix), the only precocial species measured to date. Birds were maintained for 3 days in metabolic chambers with ad lib access to food and water. A 49% increase in DEE was observed in birds laying eggs each day compared to individuals that had not yet begun laying, and DEE was positively related to daily energy deposited in the eggs. In a later study, Ward (1996) measured DEE with doubly labeled water in barn swallows Hirundo rustica that were either producing eggs, incubating or provisioning nestlings. In this case, egg production had no significant effect on DEE variation when compared to the other breeding stages and was not related to egg energy content or clutch size. In a similar study, also using doubly labeled water, Stevenson and Bryant (2000) reported contrasting results in free-living, egg-producing great tits. Controlling for the effect of ambient temperature, a significant positive relationship between DEE and egg mass was found in one year but the relationship was reversed the following year, which they reported as offering poor breeding conditions. DEE integrates energy expended in all aspects of a bird s activity budget, including behavioral adjustments to compensate for energy demanding activities such as egg production (see below, Williams and Vézina, 2001; Vézina et al., 2006; Williams et al., 2009; Salvante et al., 2010). Therefore, discrepancies in estimates of egg production costs are perhaps to be expected between studies that are using only DEE to estimate costs of egg formation. Furthermore, measures of DEE are not directly comparable to estimations based on RMR variation. Clearly, measurements on more species throughout the altricial-precocial spectrum, using a standardized approach such as measurement of resting metabolic rate during the non-breeding and egg production stages, are required before generalized interpretations of the energy costs of producing eggs can be made. Measurements of the energetic costs associated with egg production should also be undertaken in a variety of species exhibiting a wide range of body masses and egg sizes (relative to body mass) in order to examine how these parameters may affect reproductive energy investment in birds. For example, non-passerine species that lay relatively large and yolk-rich eggs (e.g. galliforms, waterfowl and seabirds) that develop over a longer time period (5 30 days, Roudybush et al., 1979; Astheimer et al., 1985; Astheimer and Grau, 1985; Astheimer and Grau, 1990, Alisauskas and Ankney, 1994; Esler, 1994; Gorman et al., 2007; Schneider, 2009) relative to small passerines species (e.g. 2.5 4 days; Ricklefs, 1974; Badyaev et al., 2005) may potentially experience different energy demands for egg formation. Laying a large clutch does not increase daily energy demand but extends total duration of the energy investment (Nager, 2006). Therefore seasonal changes in egg size within or between clutches in some species (Christians, 2002) could also reflect specific energy investment strategies adjusted to seasonal changes in local breeding conditions. Knowing that egg production in birds has a measurable energetic cost begs the question, what drives the increase in metabolic rate measured during egg formation? To date, there has been no clear answer to this question but a few studies have improved our understanding of the phenomenon. At the inter-individual level, variation in the mass of the oviduct, the organ responsible for egg albumen and shell formation, has been shown to correlate positively with RMR in egg laying house sparrows (r 2 = 0.30, Chappell et al., 1999), European starlings (r 2 = 0.18, Vézina and Williams, 2003) and zebra finches (r 2 = 0.23, Vézina and Williams, 2005a). Although variation in oviduct mass explains at best 30% of RMR variation during egg production, it is important to note that oviduct mass is also correlated with egg size, at least in European starlings (Ricklefs, 1976; Vézina and Williams, 2003). Furthermore, Christians and Williams (2001a) found that oviduct mass ex-

780 Current Zoology Vol. 56 No. 6 plains approximately 21% of the egg s albumen protein content in that same species. Taken together, these observations suggest that generating a large oviduct, despite the associated metabolic cost, may be advantageous for female birds as it may allow for laying larger, better quality eggs (Vézina and Williams, 2003), which likely improves early chick survival (Williams, 1994; Christians, 2002; Wagner and Williams, 2007). However, in addition to the added weight, maintaining a large oviduct between breeding attempts may be maladaptive due to its associated energy demands. This could explain why this organ begins to regress rapidly from the ovary to the cloaca after the last ovulation while the last follicle is still progressing down the oviduct, even though the bird may breed again in the same season (Vézina and Williams, 2003; Williams and Ames, 2004). Whether individual females can fine tune the size of their oviduct to influence egg size and quality is unknown. However, several species show an intraseasonal decline in egg size with laying date (Christians, 2002) and recent evidence suggests that females of some species can voluntarily adjust egg size to mate quality (Cunningham and Russel, 2000; Uller et al., 2005). Therefore, given the reported range of flexibility in oviduct mass (Table 1) and given the fact that its recrudescence is under hormonal control (Yu et al., 1971), it is reasonable to hypothesize that females may be able to modulate part of their physiological reproductive investment through adaptive phenotypic flexibility in oviduct size and function. More research is needed to test this oviduct flexibility hypothesis. Other aspects of egg production physiology have also been studied with regard to their possible contribution to the metabolic cost of egg formation, but no other obvious energy demanding processes have been identified. During egg formation, estrogens synthesized in the newly-grown ovary stimulate the liver to produce the two major yolk precursor macromolecules, vitellogenin (VTG) and yolk-targeted very-low-density lipoprotein (VLDLy) (Bergink et al., 1974; Deeley et al., 1975; Wallace, 1985; Walzem, 1996; Williams, 1998). Once produced, VTG and VLDLy are secreted into circulation and are then taken up by the ovary by receptor-mediated endocytosis and processed within the ovarian follicles into yolk, the nutrient and energy source for the developing avian embryo (Bernardi and Cook, 1960; Wallace, 1985; Stifani et al., 1988, reviewed in Williams, et al., 2001). Nager (2006) reviewed body composition data for females actively laying eggs and observed an emerging pattern for a larger liver in 6 out of 9 studies based on 7 species. Liver mass has also been shown to correlate positively with plasma vitellogenin in one of two years in European starlings (Christians and Williams, 1999). However, liver mass was found to vary independently of laying RMR in three consecutive years in the same species (Vézina and Williams, 2003), and liver maximal oxidative capacity was reduced in female starlings producing eggs (Vézina and Williams, 2005b). Furthermore, active yolk precursor production triggered by exogenous estrogen administration did not result in increased RMR in zebra finches (Vézina et al., 2003). Therefore, yolk precursors appear to be energetically inexpensive to produce, and this may explain why females forming eggs apparently produce yolk precursors in excess and saturate ovarian follicle receptors during rapid yolk development (Williams 2000, 2001; Williams et al., 2001). Egg production involves other processes that may contribute to the metabolic cost of egg formation which have not been investigated. One on them is yolk uptake rate by the ovarian follicles. Yolk uptake rate correlates with final yolk size (Christians and Williams, 2001b), and VTG/VLDLy-receptor mrna expression level in the ovarian follicle entering the rapid yolk development stage (i.e., F3) is positively correlated to the masses of the first laid egg and the largest developing ovarian follicle (i.e., F1) within the same clutch (Han et al., 2009). To date, the energetic cost of follicular yolk uptake rate, including biosynthesis of yolk precursor receptors, and the question of whether females can modulate egg size by reversibly varying yolk uptake rate remains to be studied. Females also have the capacity to modulate egg quality and alter offspring viability by varying yolk content as a function of constraints encountered during egg formation (e.g. hormones: Schwabl, 1993; Groothuis et al., 2005; Love et al., 2005, 2009, Carere and Balthazart, 2007; Love and Williams, 2008; antibodies: Hasselquist and Nilsson, 2009; antioxidants: Biard et al., 2009). The amount of energy invested in relation to egg composition and quality may therefore be related to an individual s immediate and future breeding effort (e.g. provisioning low quality chicks) and remains to be studied. 3.2 Males The energetic cost of gonadal development and function in males has received much less research attention than in females. Bioenergetic models based on tissue energy content suggest that the cost of testicular recrudescence is low and represents less than 2% of BMR (King, 1973; Walsberg, 1983). However, this approach

VÉZINA F, SALVANTE KG: Phenotypic flexibility in breeding birds 781 for estimating metabolic energy requirement of reproductive organs has been contested as it only considers tissue energy content and ignores synthesis, maintenance and function costs (Nilsson and Raberg, 2001; Vézina and Williams, 2002). Furthermore, actual empirical measurements using respirometry on females differ considerably from predictions generated by bioenergetic models based solely on the energy content of their reproductive tissues (Vézina and Williams, 2002). We are aware of only one study that has empirically investigated the effect of gonad size on metabolism in male birds. Chappell et al. (1999) found a significant correlation between RMR and gonad mass in reproductive male house sparrows. Difference in testis size translated into a 29% higher mass-independent RMR in males with the largest gonads (maximal gonad mass 1.5 times average mass) relative to that of males with the smallest organs (considered non-reproductive). The reported effect of gonad mass on metabolism in male house sparrows is therefore comparable to maximal energy investment measured in egg producing females (27% increase in RMR in great tits; Nilsson and Raberg, 2001). Chappell et al s (1999) finding clearly contrasts with recent data reported by Caro and Visser (2009) where testicular recrudescence triggered by artificial changes in photoperiod was not related to RMR of captive great tits maintained under two thermal regimes. Because cold treatment (8 C) induced a 20% increase in RMR but no effect on testicular growth, the authors concluded that gonad recrudescence was a process requiring relatively little energy (but see also Silverin et al. 2008 for effect of temperature on testis growth). A low energetic cost of testicular recrudescence and function is certainly compatible with observations of males growing testes weeks before females begin developing their reproductive organs (Caro et al., 2005, 2006, 2009) and maintaining mature testes for the whole breeding season (Table 2). However, in this specific case, the testes had not reached maturity by the time RMR was measured (Caro and Visser, 2009). Therefore, a possible relationship between testicular size and RMR in fully reproductive male great tits cannot be excluded and remains to be studied. Although the effect of circulating testosterone levels on BMR remains controversial (Wikelski et al., 1999; Buttemer and Astheimer, 2000; Buchanan et al., 2001; Buttemer et al., 2008) other aspects of male reproductive activity (e.g. behavioral changes including courtship singing and mate guarding) may also generate physiological changes associated with an increased RMR during this period. 4 The Impact of Reproductive Effort on Energy Budgets: Reproductive Flexibility Whether avian testicular growth and function is energetically costly or sufficiently constraining to elicit reversible physiological or behavioral adjustments to balance energy budgets remains unclear and requires further research. However, recent studies using captive zebra finches as a model system suggest that energy investment in egg production can lead to such observable changes in females. The next sections therefore focus on case studies highlighting phenotypic flexibility in relation to reproductive investment by female zebra finches. 4.1 Behavioral flexibility: The first line of defense With reported increases in RMR of 22% 27%, egg production in birds may be considered a low-cost activity when compared to total energy expended on a daily basis. Ward (1996) found no differences in average DEE of barn swallows either producing eggs, incubating or provisioning nestlings. However, as will be discussed below, comparing population average DEE in this context is misleading because it masks variation related to individual reproductive investment and compensation strategies (see Vézina et al., 2006; Williams et al., 2009). Five independent studies on captive zebra finches maintained at room temperature with ad lib access to food have shown a reduction in locomotor activity ranging from 40% to 65% during the period of egg production (Houston et al., 1995; Williams and Ternan, 1999; Vézina et al., 2006, Williams et al., 2009; Salvante et al., 2010), and the recorded decrease in activity has been shown to coincide with the onset of rapid yolk development in females (Williams and Ternan, 1999; Vézina et al., 2006). This phenomenon is not limited to zebra finches, however, as reduced activity has also been reported in free-living willow flycatchers Empiodonax traillii during egg formation (Ettinger and King, 1980), and female ruddy ducks Oxyura jamaicensis have been found to increase resting time during the same period (Tome, 1991). The effect of this behavioral adjustment on a female s DEE (measured by doubly labeled water) has been investigated twice over two successive breeding attempts within the same females (Vézina et al., 2006; Williams et al., 2009). In both studies, reduced locomotor activity

782 Current Zoology Vol. 56 No. 6 was interpreted as a behavioral strategy to compensate for the energy costs associated with egg production. As found by Ward (1996) in barn swallows, average DEE of female zebra finches did not change between non-breeding and egg-laying stages (Vézina et al., 2006; Williams et al., 2009). However, there was marked inter-individual variation in the initial egg production investment (measured as RMR variation in females during egg laying), and the changes in locomotor activity were negatively correlated with egg producing females RMR (Vézina et al., 2006). In other words, the level of energy compensation was adjusted to the individual investment, as females investing the most in egg production were also showing the largest reduction in activity. This latter finding paralleled a previous report of a negative relationship between clutch size and locomotor activity in laying zebra finches (Williams and Ternan, 1999). Behavioral flexibility therefore allowed birds to compensate for egg production costs and, perhaps not surprisingly, no changes or even decreases in average energy intake (measured as food consumption) have been detected in association with egg formation in this species under favorable conditions (Houston et al., 1995; Williams and Ternan, 1999, Vézina et al., 2006; Salvante et al., 2010). Consequently, balancing the energy budget through behavioral flexibility was suggested as a means to avoid large variation in DEE in females actively producing eggs (Vézina et al., 2006). Vézina et al. (2006) and Williams et al. (2009) also showed that behavioral energy reallocation was individually variable and led to a range of effects on DEE, from negative changes from the non-breeding to the egg-laying stage (i.e. overcompensation) in some individuals, to additive effects (i.e. net increase in DEE) despite behavioral energy savings, in others. Both individual investment, measured by intra-individual variation in RMR, and compensation strategy (i.e. effect on DEE) were shown to be repeatable between breeding attempts (Vézina and Williams, 2005a; Williams et al., 2009), a finding which suggests that individual females tend to maintain their reproductive effort from one breeding attempt to the next. However, the link between consistency in energy investment and reproductive output is not as clear. Although, as in many species (reviewed by Christians, 2002), clutch size and egg size were repeatable among individual females (Williams et al., 2009), a positive relationship between egg-laying DEE and clutch size was found during the first breeding attempt (Vézina and Williams, 2005a) but was not confirmed in the following attempt (Williams et al., 2009). Behavioral flexibility seems to be relatively common in birds facing energy constraints. For example, in addition to reproductive costs, reduced activity has been observed in birds facing thermoregulatory challenges (e.g. Cherel et al., 1988; Salvante et al., 2010), experiencing food limitation (e.g. Meijer et al., 1996; Dall and Witter, 1998) and undergoing feather molt (e.g. Austin and Fredrickson, 1987; Robin et al., 1989). This may be the simplest and fastest way to decrease energy expenditure and may represent the first line of defense when facing energy constraints before engaging in, perhaps slower and more costly, reversible changes in physiological function. However, balancing energy budgets through reductions in locomotor activity might not always be a feasible option in natural conditions. This may be especially true in female birds wherein gonadal recrudescence and egg formation take place before the seasonal peak of food abundance (Perrins, 1996) and likely require increased foraging activity. In such cases, what other mechanisms are available for females to balance their energy budget? Several recent studies investigated energy compensation strategies in captive birds by experimentally increasing work for food rewards, thus preventing energy savings through reduced activity. A common finding was phenotypic flexibility at the physiological level as birds typically reduced nighttime resting metabolic rate (Bautista et al., 1998; Deerenberg et al., 1998; Nudds and Bryant, 2001; Wiersma and Verhulst, 2005; Wiersma et al., 2005). 4.2 Physiological flexibility: When behavioral adjustments are not enough Although one can easily measure behavioral flexibility in response to the energetic costs of reproductive investment, perhaps more challenging is to study how birds can use phenotypic flexibility at the physiological level to adjust components of reproductive investment itself. Two recent studies (Salvante et al., 2007, 2010) suggest that female zebra finches facing conflicting constraints during egg production make physiological compromises in an attempt to maintain reproductive output. Salvante et al. (2007) combined the energetic cost of producing eggs with increased thermoregulatory requirements. Using a repeated measures design, zebra finches were acclimated and bred at both 7 C and 21 C. Birds increased food intake by 45% in the cold as a result of higher daily energy demands but no changes in body condition, egg mass or mass of egg components (i.e., yolk, albumen and shell) were recorded. Ambient